[vector] RFR: Remove VShiftV nodes

[Vladimir Ivanov]

Thanks, Sandhya.

> I meant to say the following version of the method can be moved to x86.ad:
> 1581 bool Matcher::supports_vector_variable_shifts(void) {
> 1582   return (UseAVX >= 2);
> 1583 }

I haven't decided how I prefer to handle x86-32.

The port lacks proper testing, so it would be nice to keep it intact.

On the other hand, I don't want to put everything into x86_64.ad / under 
#ifdef _LP64.

So, let's leave it as it is for now.

Best regards,
Vladimir Ivanov

> -----Original Message-----
> From: Viswanathan, Sandhya
> Sent: Thursday, April 09, 2020 4:08 PM
> To: Vladimir Ivanov <vladimir.x.ivanov at oracle.com>; panama-dev <panama-dev at openjdk.java.net>
> Subject: RE: [vector] RFR: Remove VShiftV nodes
> 
> Hi Vladimir,
> 
> I have only one comment, the following method can be in x86.ad:
> 
> 1374 bool Matcher::supports_vector_variable_shifts(void) {
> 1375   return false; // UseAVX >= 2
> 1376 }
> 
> Both 32-bit and 64-bit support vector_variable_shifts when UseAVX >= 2.
> 
> Overall, the code looks much cleaner now.  As John said, the code is becoming smaller and smaller :).
> 
> Best Regards,
> Sandhya
> 
> -----Original Message-----
> From: panama-dev <panama-dev-bounces at openjdk.java.net> On Behalf Of Vladimir Ivanov
> Sent: Thursday, April 09, 2020 1:20 PM
> To: panama-dev <panama-dev at openjdk.java.net>
> Subject: [vector] RFR: Remove VShiftV nodes
> 
> http://cr.openjdk.java.net/~vlivanov/panama/vector/vshift/webrev.00/
> 
> Get rid of dedicated nodes for variable shifts and reuse existing [UR|R|L]ShiftV nodes.
> 
> Overall, it simplifies the code except detecting when variable shifts are supported. Introduced Matcher::supports_vector_variable_shifts() for that.
> 
> The patch is against vectorIntrinsics branch.
> 
> Best regards,
> Vladimir Ivanov
>